From: Jeffrey Kay (jkay@ENGENIA.COM)
Date: Thu Dec 21 2000 - 07:09:42 PST
Interesting article indicating that P2P technology lacks serious business
The Defogger: Beware the Next Tech Craze: P2P
Issue: December 2000
Print Article | Email This Article
Despite the buzz, peer-to-peer startups lack business benefits.
---- Over the past few months, you've heard a lot about a new tech craze called peer-to-peer computing. Brace yourself, because you're going to hear a lot more if the dollars that VCs are throwing at the new field are any indication. Then there is the alien tie-in. But more on that later.
Generally speaking, peer-to-peer refers to a system that allows a user on a network to draw on the spare processing power and/or storage capacity of other computers (peers) on the same network. It's an attractive concept, given that most of the world's desktop PCs are idle much of the time and have plenty of surplus storage space on their hard drives. While you're getting a cup of coffee or sleeping at night, your computer is essentially doing nothing. Wouldn't it be great if we could put some of that unused processing power to work -- for example, to speed up your Net access? Or maybe to detect evidence of intelligent alien life-forms?
The latter is the idea behind University of California at Berkeley research project SETI@home, which kicked off the recent P2P craze. Since mid-1999, SETI@home has been using a form of P2P computing to analyze interstellar radio signals that may contain evidence of some alien civilization. (SETI is an acronym for the Search for Extraterrestrial Intelligence.) The software takes advantage of participating computers' idle time to run complex computations. While it works, a screen saver that comes with the program displays impressive graphics that reflect the radio wave analysis in action -- real gee-whiz stuff. After your computer crunches a small piece of the puzzle, it sends the results back to SETI@home headquarters in Berkeley.
Altogether, more than 2.4 million people have donated their unused processing cycles to the SETI@home project. The combined output of those machines varies from day to day but usually tops the performance of the fastest single computer in the world -- a $110 million IBM supercomputer used by the U.S. Department of Energy to simulate nuclear weapons tests.
SETI@home's success at finding users (if not aliens) has convinced a lot of entrepreneurs that P2P has real potential. SETI@home, Napster, Gnutella, and other headline-grabbing P2P technologies are merely the first wave. "P2P will be the B2B of 2001," gushes Mark Eggleston, an emerging technologies analyst at Currid and Co., a Houston-based consulting and research firm.
Venture capitalists, hoping to cash in on the opportunity, are writing fat checks to P2P startups. For example, Groove Networks, which in October unveiled a way for teams to share documents and applications using a P2P approach, has pulled in $60 million in funding, led by Accel Partners. United Devices, which aims to harness SETI@home-style distributed processing for commercial ends, bagged $13 million from Softbank Venture Capital and Oak Investment Partners. And Napster, the P2P network that needs no introduction, landed $15 million last spring from A-list VC Hummer Winblad.
Even staid Intel is getting into the game with its Peer to Peer Working Group, a sort of clearinghouse for information on P2P ideas and an aspiring standards-setting body. Intel claims that more than 100 P2P companies are seeking membership. (Intel sees P2P as a good thing, since it puts Intel Pentium processors to work more often, which it hopes will increase the demand for PCs.) "We believe that this is a new inflection point in the worldwide Internet economy," says Bob Knighten, Intel's P2P evangelist and chairman of the working group. "We don't know exactly where peer-to-peer will win but are confident that it will win big."
But hold on a second. Is P2P for real, or just another overfinanced tech bubble in the making? When you look hard at P2P, it's difficult to get as fired up as the evangelists. SETI@home is a strictly noncommercial venture, and people donate their CPU time because, well, it's cool to help hunt for aliens. Napster, currently embroiled in a major copyright infringement lawsuit, has yet to make a dime -- or even say how it plans to do so. AOL's and Universal's plans to offer all-you-can-eat digital music download services to paying subscribers don't help its chances. (For more on Napster's legal battle, see "Can Napster Beat the Rap?," page 69.)
And there are even some signs that P2P doesn't work all that well. A paper released in September by Xerox PARC researchers Eytan Adar and Bernardo Huberman argues that Gnutella (a P2P file-sharing technology initially developed by AOL subsidiary Nullsoft) is ill-served by its utopian premise that people will share their resources freely and equally. Seventy percent of Gnutella users, according to the study, are freeloaders who download files from the network but refuse to give back by sharing files of their own. You can't build a big peer-to-peer network unless you have a lot of peers contributing resources (storage, processing, or whatever).
Ultimately, where P2P will work as a business, if it works at all, is in more controlled circumstances, where there are clear benefits for participating in the network and a clear business model.
United Devices, based in Austin, Texas, is hoping to turn SETI@home's P2P model into a moneymaking venture for itself that also rewards those who rent out time on their PCs to the company. Volunteers run the United Devices "agent" program on their PCs, and the agent quietly makes use of the computers' idle time to perform calculations. United Devices sells this combined processing power to companies with computation-intensive needs, such as genome research or data-mining. United Devices started testing its network in late September, and signed up "thousands of users" in its first two weeks, says company CEO Ed Hubbard. To encourage volunteers, United Devices hosts sweepstakes and gives away "webmiles" (which can be exchanged for airline tickets). If this sounds suspiciously similar to "earn while you surf the Web" ploys, you're right. And if you recall that the dotcoms that tried that business are either dead or dying, you're right about that too.
Another idea is to use a P2P model to help your company reduce congestion on its Internet connection. MangoSoft, a software developer based in Westborough, Mass., makes a product called Cachelink that uses P2P file sharing to increase the efficiency of Web browsing. Ordinarily, when you look at a webpage, your browser stores the page in a cache file on your hard drive. The next time you visit that site, the browser can quickly display the cached version instead of downloading the same file all over again. Cachelink extends this idea into the P2P world by connecting the browser caches of all users over a local network. If I visit a webpage that someone else in my company has already looked at, Cachelink will send me the page from the network cache rather than downloading it from the Net. That's faster because it uses the speedy local area network instead of a slower connection to the public Internet. It's also cheaper.
"In an enterprise where you're using a lot of the same material and going to a lot of the same websites, you can save a lot of bandwidth, ISP costs, and time," says Michael Hoch, an analyst with Boston-based Aberdeen Group. Cachelink, which has been available since August 1999, charges about $20 per user. With plummeting bandwidth costs, however, it's not clear how valuable this caching will be in the long run.
Groove Networks, led by Lotus Notes creator Ray Ozzie, thinks P2P is a good approach to helping remote team members work together on projects over a network. When you run the Groove software, you can create secure, shared virtual spaces and invite others to join you. Think The Matrix, only you aren't there physically. In these spaces, everyone on a team can simultaneously work on a whiteboard, a textual report, an architectural drawing, or any other kind of file. The data is stored not on a central server but on each participant's hard drive.
All you need to participate or to create a new, private workspace is the free Groove software, which is a sizable 10MB download. Ozzie says Groove, which launched in beta on Oct. 24, will make money by selling customizable versions of the software to companies and to software developers, who will link the software to existing intranet applications. But Groove is up against competitors like Microsoft's NetMeeting and WebEx that offer similar ser-vices, just not in the P2P model.
Whither peer-to-peer? We know that where P2P has drawn users, there is no known business model to go along with it. Where there may just be a real business, as could be the case with Groove Networks and some other business-oriented startups, it's not clear that a strong need exists for what they aim to provide. Call me a skeptic, but for now, this P2P bandwagon is one that I'll let pass by.
To learn more, go to www.ecompany.com and type P2P in the search box to find the Web File, a collection of links and additional information on this story.
Read Dylan Tweney's new column every Thursday at www.ecompany.com.
Jeffrey Kay <email@example.com> Chief Technology Officer, Engenia Software Inc.
"First get your facts, then you can distort them at your leisure" -- Mark Twain "Golf is an endless series of tragedies obscured by the occasional miracle" -- Sports Illustrated "If A equals success, then the formula is A equals X plus Y plus Z. X is work. Y is play. Z is keep your mouth shut." -- Albert Einstein
This archive was generated by hypermail 2b29 : Thu Dec 21 2000 - 07:15:23 PST